No Stata talk would be complete without the auto
dataset.
. sysuse auto
(1978 Automobile Data)
We can make a better term for energy usage
. ** have: distance per energy; want: energy per distance . gen gp100m = 100/mpg . label var gp100m "Gallons per 100 miles"
We can do regressions...
. regress gp100m weight c.displacement##foreign
Source | SS df MS Number of obs = 74 -------------+------------------------------ F( 4, 69) = 64.16 Model | 94.2376637 4 23.5594159 Prob > F = 0.0000 Residual | 25.3385971 69 .367226045 R-squared = 0.7881 -------------+------------------------------ Adj R-squared = 0.7758 Total | 119.576261 73 1.63803097 Root MSE = .60599 ------------------------------------------------------------------------------ gp100m | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- weight | .0012254 .0002216 5.53 0.000 .0007832 .0016676 displacement | .0029178 .0018215 1.60 0.114 -.0007159 .0065516 | foreign | Domestic | 0 (base) Foreign | -1.161456 .708529 -1.64 0.106 -2.574933 .2520208 | foreign#| c. | displacement | Foreign | .0156473 .005816 2.69 0.009 .0040447 .02725 | _cons | .5714418 .4522969 1.26 0.211 -.3308659 1.47375 ------------------------------------------------------------------------------
And we can put results directly into the running text. For instance, we can see that the the coefficient for weight is 0.00123 and that it has a p-value of 0.0000005.
Embedding graphs requires using a piece of free software (ImageMagick). This will need some massaging to make it easy enough for everyone to use.